The Study of Distributed Fault-Tolerant K-Mutex Algorithm

• Main Author: 劉康佑
• Other Authors: 葉耀明
• Format: Others
• Language: zh-TW
• Published: 1999
• Online Access: http://ndltd.ncl.edu.tw/handle/14971296765767183785

碩士 === 國立臺灣師範大學 === 資訊教育研究所 === 87 === In the distributed system, network resources shared by many nodes results in the mutual exclusion problem of the shared resources. Mutual exclusion require that a resource can only be assignned to a node at a time. If the system has one shared source, it is called 1-mutex problem. When a system has k shared sources, it is called k-mutex problem. Mutual exclusion is crucial for the design of a distributed system. Many operations such as data replication and distributed memory sharing all require the mutual exclusion mechanism to solve the coordination of resource sharing. In this thesis, we devise two new distributed algorithms to achieve mutual exclusion. The first method is called PMQ algorithm. It is a permission-based algorithm, which not only reduces the communication cost using coterie, but also provides high fault-tolerant capability. The second method is called Two_Token algorithm. It is a Token-based algorithm, which uses two tokens (one is A_Token; the other is C_Token) to provide the fault tolerant capability. Both algorithms can solve 1-mutex and K-mutex problem, and provide better fault-tolerant capability than the existing algorithm.